“I agree to them.” Four words can appear conclusive in the record of a criminal trial. But their meaning depends on what “them” actually refers to: which document, which version, and precisely which facts the agreement covers.

In my case, those words became part of the explanation for why I had waived my right to confront certain witnesses. I was convicted in 2024 in federal court in Texas on 33 counts of aiding the preparation of false and fraudulent tax returns and sentenced to 188 months in prison. The Fifth Circuit affirmed my convictions and sentence. I dispute that outcome, but my disagreement does not make my account an established finding.

The broader question is how an institution verifies the materials on which its judgments depend. A stipulation records an agreement. A transcript records what was said. Disputes over either can influence what an appellate court has available to review, although they raise different legal and factual questions.

Stipulations can spare witnesses unnecessary travel and narrow a trial to genuinely contested issues. They may concern a document’s authenticity, what a witness would testify to, or a fact both parties accept. These are fundamentally different commitments. Understanding precisely what was agreed to matters as much as establishing that someone agreed.

I maintain that my email response concerned proposed stipulations that differed from the substantive document ultimately filed. I have also challenged the authorization for signatures attributed to the defense. These remain disputed contentions, not established findings of forgery or invalidity.

The government’s evidence, however, extends beyond my four-word response. The filed email chain includes one defense lawyer writing, “We believe that yours are acceptable.” Another lawyer forwarded my response to the prosecutor with two named attachments and wrote, “We agree. Done!” Any account that omits these communications would leave out evidence supporting the government’s position.

In its August 11 opinion, the Fifth Circuit relied on the email exchange and defense counsel’s conduct during the trial. It concluded that the stipulations reflected a legitimate trial strategy and that I had waived my confrontation rights. The court further explained that, under its precedents, defense counsel can make such a waiver without the defendant’s separate affirmative agreement, provided the defendant does not dissent and the decision represents a legitimate trial tactic.

That ruling is essential to understanding the dispute. But so is the distinction between identifying an agreement and determining its legal consequences. An original email containing its actual attachments, the document ultimately filed, and the authorization associated with that document can help establish whether the same text remained consistent throughout the process. A flattened email printout may leave questions about attachments unanswered. Neither a missing attachment indicator nor the absence of a handwritten signature, by itself, establishes wrongdoing.

A dispute over a trial transcript presents a different challenge: determining whether the written record faithfully captures the words actually spoken.

During my trial, a witness was questioned about his wife’s work-related vehicle expenses. The certified transcript records his response as: “I would think — I would say, no.” In a filed motion, I asserted that I heard: “I would think so. I wouldn’t say no.”

Those statements convey substantially different meanings. Yet the second version rests on my recollection. I have not listened to an underlying recording that independently verifies it. Even if my proposed correction were established, that would not, by itself, prove the deduction was legally permissible or require my convictions to be overturned.

The district court denied my transcript-correction and source-access motions in June 2025, finding that I had not presented sufficient evidence to overcome the certified transcript’s presumed accuracy. The court also criticized language in my motions suggesting that an audio comparison had occurred, despite my stated inability to obtain the recording. Its order did not describe any comparison of the disputed passages against the original reporting materials.

The Fifth Circuit did not address the merits of that transcript ruling in its August decision. Instead, it dismissed that portion of the appeal after determining that a separate notice of appeal was required.

This is an important distinction. An appellate court’s refusal to consider an issue does not establish that the disputed words were transcribed correctly. Nor does it establish that they were recorded incorrectly.

Federal law already provides mechanisms for challenging an official transcript. The Court Reporter Act gives certified transcripts a presumption of correctness while separately providing for the preservation and inspection of original reporting records. Federal Rule of Appellate Procedure 10(e) assigns district courts responsibility for resolving disputes over whether the record accurately reflects what occurred.

Neither provision guarantees a new hearing whenever a litigant remembers events differently. Nor does federal law require every proceeding to be accompanied by an audio recording. In my case, the district court treated any backup audio as the court reporter’s private property. But backup audio and official original reporting notes are not necessarily interchangeable. Their distinct legal status matters when determining what materials may be available for verification.

These distinctions suggest several procedural improvements worth considering without presuming misconduct.

For stipulations, a version-specific confirmation could identify the precise text being accepted while distinguishing admitted facts from agreed descriptions of anticipated testimony. Documenting when and for what purpose a court accepted an agreement would make its subsequent use easier to trace. A more extensive personal inquiry could clarify a defendant’s understanding, although requiring one for every routine evidentiary agreement would consume time and potentially complicate ordinary trial management.

For transcripts, a focused verification request could identify the precise passage in dispute, the proposed alternative wording, why the difference matters, and the evidence supporting the challenge. Where further examination is warranted, comparing available source materials under court supervision is fundamentally different from publicly releasing an entire recording.

Protected information, incomplete recordings, and unavailable source materials would each present separate complications. A challenge can be specific without being substantiated, and specificity alone does not eliminate the need to evaluate the available evidence.

A reasoned judicial determination could then distinguish among several possible outcomes: the source material supports the certified transcript; it supports a correction; it cannot resolve the discrepancy; or the evidence presented does not justify further inquiry. Explaining which conclusion was reached would make the decision more transparent without predetermining the result.

Verification and legal relief must also remain separate questions. Establishing an error in the record does not automatically determine whether that error affected the judgment, whether an objection was properly preserved, or which legal procedure might provide a remedy. Conversely, a procedural barrier to relief does not resolve the underlying historical question of what was actually said.

In an earlier article, I examined how individuals might challenge inaccuracies in AI-generated biographies. A disputed court record brings that same problem closer to the source of institutional authority. The question is not whether my preferred account should replace the official version simply because I contest it. The question is whether a meaningful process exists to test the difference, examine the evidence, and explain what the record can actually establish.

When a judicial decision depends on an official record, the integrity of that record matters. So does the ability to distinguish between what has been verified, what remains disputed, and what the courts have declined to decide.